In some views, a major driving force of learning is the propensity to apply whatever understanding one may have to novel situations. However, controversy has always surrounded the topic of transfer of knowledge. Extravagant claims that transfer is omnipresent or, antithetically, that it rarely occurs come from two different historical traditions, problem solving and conceptual development. The proposed research will draw upon both traditions in an attempt to understand transfer in very young children (1-5 years). In order to evaluate the likelihood that transfer will occur, it is necessary to classify the knowledge to be transferred in terms of its structural organization and functional significance to the learner. If knowledge is consistent with a coherent theory or a principled understanding, it is nearly impossible to impede flexible application. If, however, application of a previously learned isolated rule or specific solution is required, transfer is elusive. Learners are hesitant to apply fragmentary, unassimilated knowledge which tends to remain embedded within the specific contexts of habitual use. It is argued that this pattern is age-independent. What changes with age is the complexity and organization of available knowledge. In this research, children's application of specific solutions (tool use/problem solving) or general principles (biological knowledge/conceptual development) will be compared and contrasted. A model of transfer is proposed that involves three methods of fixing or encapsulating information, hence impeding transfer, and three methods of releasing or disembedding a solution so that it becomes widely available. Encapsulating mechanisms are: (1) negative learning set; (2) cognitive embeddedness; and (3) functional fixedness. Freeing mechanisms are: (1) positive learning set; (2) perceiving the solution tool as one of many creative uses; and (3) reflecting on the deep structure via teaching or discussion.